PROJECT SUMMARY Effective, affordable, and scalable weight loss (WL) interventions are needed to reduce obesity prevalence rates and improve obesity-related co-morbidities. Behavioral WL programs delivered via the Internet offer a low-cost alternative to traditional face-to-face programs and have high dissemination potential, as the intervention can be delivered entirely remotely. While these standardized Internet programs are effective for many, a significant proportion of individuals still fail to achieve a clinically significant WL and may require a more personalized treatment approach. One strategy for improving Internet-delivered WL (IDWL) treatment, while still limiting delivery costs, is to use a stepped-care intervention approach, in which all individuals receive a standardized Internet program and a more intensive intervention is reserved for those with suboptimal WL outcomes. Previous research indicates that 4-week WL is strongly associated with long-term WL; thus 4 weeks appears to be an opportune time to identify and provide early rescue efforts to initially non-responsive individuals. Our research team has demonstrated that the provision of a brief period of individualized coaching to early non-responders enrolled in IDWL treatment significantly improved intervention adherence and post- treatment WL, when compared to early non-responders who received no coaching. The current study builds on these findings and examines whether WL outcomes can be further improved through the use of extended coaching, whether these effects are sustained once the phone coaching (PC) is removed, and whether these interventions are effective from a cost perspective. 450 individuals will be randomized to: 1) IDWL+Brief PC for early non-responders (`Brief'), 2) IDWL+Extended PC for early non-responders (`Extended'), or 3) IDWL only (`Control'). All participants will receive an automated 4-month IDWL program and an 8-month IDWL maintenance program. Early non-responders (4-wk WL <4%) randomized to `Brief' will receive 3 individual phone calls, while those randomized to `Extended' will receive weekly phone calls throughout the WL program (12 in total). Early non-responders randomized to `Control' will receive no PC. Assessments of weight, diet, physical activity, and weight-related behaviors will be performed at baseline, 4, and 12 months. The primary aim is to compare `Brief', `Extended' and `Control' on WL at 4 and 12 months. This study will also assess the effect of moderators (sex, race/ethnicity, and early non-responder subgroup: 4-wk WL <2% vs. 2-4%) and mediators (diet, physical activity, intervention adherence) on weight change, as well as the cost/kg of WL across the 3 treatment arms. Findings from this study can be used to develop algorithms for choosing the optimal amount of PC for early non-responders enrolled in an IDWL program. If effective, this program could provide a low cost, yet personalized treatment approach for `at-risk' individuals receiving IDWL treatment. Internet programs which incorporate PC for early non-responders could then be translated into healthcare and community settings.